IFNGR1 signaling is associated with adverse pregnancy outcomes during infection with malaria parasites

Complicated/severe cases of placental pathology due to Plasmodium falciparum and P. vivax, especially adverse pregnancy outcomes during P. vivax infection, have been increasing in recent years. However, the pathogenesis of placental pathology during severe malaria is poorly understood, while responses against IFN-γ are thought to be associated with adverse pregnancy outcomes. In the present study, we explored the role of IFN-γ receptor 1 (IFNGR1) signaling in placental pathology during severe malaria using luciferase-expressing rodent malaria parasites, P. berghei NK65 (PbNK65L). We detected luciferase activities in the lung, spleen, adipose tissue, and placenta in pregnant mice, suggesting that infected erythrocytes could accumulate in various organs during infection. Importantly, we found that fetal mortality in IFNGR1-deficient mice infected with PbNK65L parasites was much less than in infected wild type (WT) mice. Placental pathology was also improved in IFNGR1-deficient mice. In contrast, bioluminescence imaging showed that parasite accumulation in the placentas of IFNGR1-deficient pregnant mice was comparable to that in WT mice infected with PbNK65L. These findings suggest that IFNGR1 signaling plays a pivotal role in placental pathology and subsequent adverse pregnancy outcomes during severe malaria. Our findings may increase our understanding of how disease aggravation occurs during malaria during pregnancy.


Introduction
Malaria is a devastating parasitic disease in tropical and subtropical regions, with an estimated 214 million cases and 438,000 deaths per year [1]. The populations at greatest risk of developing severe pathology are children under the age of 5 years and pregnant women in areas where Plasmodium falciparum is endemic. Placental malaria is characterized by the accumulation of infected erythrocytes and inflammatory cells in the placenta [2,3]. Placental malaria has been reported to be correlated with adverse pregnancy outcomes such as fetal growth restriction, still birth, premature delivery and, possibly, preeclampsia [4,5]. PLOS  for 1 day with a male B6 mouse aged > 9 weeks and examined for the presence of a vaginal plug the next morning. Mice with or without a vaginal plug were infected with malaria parasites on day 12 post-mating.

DNA constructs
The SK-1 construct contained a selection cassette consisting of green fluorescent protein gene (gfp) and a pyrimethamine resistance gene, human dihydrofolate reductase-thymidylate synthase (hdhfr) [33]. The expression of gfp and hdhfr is controlled by hsp70 (PBANKA_ 071190) and elongation factor-1 (PBANKA_113340) promoters, respectively. Plasmid containing luciferase gene (pLG4. 10[luc2]) was purchased from Promega (Madison, WI, USA). To generate the luciferase-expressing cassette, luciferase gene (luc2) in the plasmid was amplified by PCR using specific primers (S1 Table). The PCR product of luc2 was cleaved using the NheI and BglII restriction enzymes, and gfp of SK-1 was replaced with luc2. Luciferase-expressing cassette was introduced into the ORF of the targeted gene by double-crossover homologous recombination [34]. The gene-targeting vector was prepared by PCR [35]. Briefly, the 5 0 and 3 0 regions flanking the ORF of the target genes, p230 [36], were amplified by PCR. The PCR products were annealed to either side of the luciferase-expressing cassette and amplified by PCR using gene-specific primers (S1 Table).

Parasites and infections
Pb NK65 is a lethal strain and was originally obtained from Dr. M. Yoeli (New York University Medical Center, New York, NY, USA). Infected erythrocytes of Pb NK65 parasites were cultured for 18 h under standardized in vitro culture conditions. Mature schizonts were then collected by Nycodenz density-gradient centrifugation [34]. Transformations were performed using the Amaxa Basic Parasite Nucleofector Kit (Amaxa GmbH, Cologne, Germany) according to the manufacturer's protocol. Briefly, 5 × 10 6 to 5 × 10 7 purified Pb NK65 mature schizonts were mixed with 100 μL of Nucleofector solution containing 5 μg of gene-targeting vector. Transfections were then completed using the Amaxa Nucleofector electroporation program U-33. Transfected parasites were injected intravenously (i.v.) into naïve B6 recipient mice. At 30 h post-injection, transfected parasites were isolated by the addition of pyrimethamine to the drinking water of infected mice. After parasitemia returned to detectable levels post-selection, transfected parasites were cloned by limiting dilution, after which a single parasite was injected into a mouse to ensure a clonally pure population. Cloned transfected parasites were stored as frozen stocks in liquid nitrogen. Infected erythrocytes of transfected parasites were generated in donor mice inoculated intraperitoneally with each frozen stock of parasite. The donor mice were monitored for parasitemia daily and bled for experimental infection in ascending periods of parasitemia. Experimental mice were infected intravenously with 1 × 10 4 infected erythrocytes or 5 × 10 6 to 5 × 10 7 purified mature schizonts of a given parasite strain.

Parasitemia and hematocrit
Blood was observed by microscopic examination of methanol-fixed tail blood smears stained with 3% Giemsa diluted with phosphate buffer, pH 7.2, for 45 min. The number of infected erythrocytes in 250 erythrocytes was enumerated when parasitemia exceeded 10%, whereas 1 × 10 4 erythrocytes were examined when mice showed lower parasitemia. The percentage of parasitemia was calculated as follows: [(number of infected erythrocytes)/(total number of erythrocytes)] × 100.
For hematocrit measurement, blood was obtained from pregnant uninfected and pregnant infected mice on days 5 and 7 post-infection. The blood (50 μL) was collected into a heparinized capillary tube and centrifuged at 12,000 × g for 5 min with a micro-hematocrit centrifuge (HC-12A; Tomy, Tokyo, Japan). Hematocrit was expressed as the percentage of blood cells in the total volume of blood.

Ex vivo organ bioluminescent imaging
Bioluminescent imaging was performed with a Photon IMAGER system (Biospace Lab, Nesles la Vallée, France). Mice were anesthetized and administered 1 mg of VivoGlo™ Luciferin (In Vivo Grade) dissolved in 150 μL of phosphate buffered saline (PBS) by i.v. injection. After receiving the VivoGlo™ Luciferin, mice were killed and organs were collected for image acquisition. Acquisition of emitted photons, with a charge-coupled device camera, was monitored. Ex vivo bioluminescent imaging data were analyzed using the M3software (Biospace) with size-constant regions of interest (ROIs).

Histological examination of placentas
Placentas were obtained from uninfected pregnant mice and infected pregnant mice on day 6 post-infection (p.i.). Mice were killed and placentas were removed. The placentas were fixed in 10% buffered formalin and embedded in paraffin. Six-micrometer-thick sections were stained with hematoxylin and eosin (H&E). The stained thick sections were photographed at 20×, 100×, and 400× magnification using an All-in-One Fluorescence Microscope (BZ9000; KEY-ENCE Japan, Osaka, Japan).

Statistical analysis
Parasitemia, hematocrit, luciferase activity, and levels of cytokines were analyzed using the Student's t-test, which was performed using Statcel (OMS, Saitama, Japan). P values of less than 0.05 were considered statistically significant.

Pregnant mice show severe pathology and adverse pregnancy outcomes during infection with malaria parasites
We first generated luciferase-expressing Plasmodium berghei NK65 (PbNK65L) (S1 Fig) and examined the outcomes of infection with PbNK65L. As shown in Fig 1A, parasitemia of pregnant mice infected with PbNK65L was rapidly increased compared with nonpregnant mice from day 3 p.i. In pregnant mice infected with PbNK65L, the levels of hematocrit were comparable to those in infected nonpregnant mice until day 5 p.i. (Fig 1B). However, infected pregnant mice showed lower levels of hematocrit than those in infected nonpregnant mice on day 7 p.i. (Fig 1B). Preterm delivery was observed in pregnant mice infected with PbNK65L and the survival rate of pups was decreased compared with that of uninfected mice (Table 1). These results suggest that pregnant mice show severe pathology compared with nonpregnant mice and adverse pregnancy outcomes during infection with malaria parasites.
To examine whether preterm delivery and fetal death occur during infection with malaria parasites, pregnant mice on day 12 or 15 post-mating were infected with intact PbNK65L (S2 Table). In pregnant mice on day 12 post-mating, preterm delivery and fetal death were observed during infection (S2 Table). In contrast, pregnant mice on day 15 post-mating successfully delivered live pups during infection (S2 Table). Since the pregnancy period was 18 days in mice that were infected with PbNK65L on day 12 post-mating, the risk of preterm delivery appears to increase from day 6 p.i. Several studies reported that production of proinflammatory cytokines, such as IFN-γ and TNF-α, was enhanced in placenta from pregnant women or mice infected with malaria parasites [15-19, 30, 31]. To examine whether the response of pro-inflammatory cytokines is involved in adverse pregnancy outcomes, we first measured levels of cytokines in plasma on day 5 p.i. The high levels of IFN-γ were observed in infected pregnant mice compared with those in infected nonpregnant mice on day 5 p.i. (Fig  1C). Analyses of cytokines in plasma detected high levels of IL-10 in uninfected and infected pregnant mice on day 5 p.i. (Fig 1D). These findings suggest that the immune response activated by IFN-γ plays an important role in adverse pregnancy outcomes during PbNK65L infection.

Pregnant IFNGR1-KO mice infected with malaria parasites successfully delivered live pups
To investigate whether the immune response to IFN-γ is involved in adverse pregnancy outcomes during malaria infection, pregnant IFN-γ receptor 1-deficient (IFNGR1-KO) mice were  infected with PbNK65L. As shown in Fig 2A, parasitemia in pregnant and nonpregnant mice infected with PbNK65L was not affected by deficiency of IFNGR1. However, the survival rate of pups of pregnant IFNGR1-KO mice was higher than that of wild-type (WT) mice during infection (Table 1). Preterm delivery, which was observed in infected pregnant WT mice on day 6 p.i., did not occur in pregnant IFNGR1-KO mice (Table 1). Therefore, we next examined the hematocrit and IFN-γ and IL-10 levels in pregnant WT and IFNGR1-KO mice on day 6 p. i. In pregnant IFNGR1-KO mice infected with PbNK65L, the hematocrit was comparable to that in infected pregnant WT mice on day 6 p.i. (Fig 2B). High IFN-γ levels were observed in infected pregnant IFNGR1-KO mice compared to those in infected pregnant WT mice on day 6 p.i. (Fig 2C), while the IL-10 levels in infected pregnant IFNGR1-KO mice were comparable to those in infected pregnant WT mice on day 6 p.i. (Fig 2D). Erythrocytes infected with PbNK65L accumulate in the lung, spleen, adipose tissue, and placenta in pregnant IFNGR1-KO mice Next, we investigated the organs in which PbNK65L-infected erythrocytes accumulate in pregnant IFNGR1-KO mice using bioluminescence imaging. On day 3 p.i., luciferase activities were detected in the lung, liver, spleen, subcutaneous adipose tissue, and placenta of infected pregnant WT and IFNGR1-KO mice (Fig 3A). However, the luciferase activities in the liver and placenta were much lower than those in the lung, spleen, and subcutaneous adipose tissue (Fig 3A). By contrast, in infected pregnant WT and IFNGR1-KO mice, the luciferase activity in subcutaneous adipose tissue was higher than in infected nonpregnant WT and IFNGR1-KO mice (Fig 3A and 3B). On day 6 p.i., high luciferase activity was detected in the lung, spleen, subcutaneous adipose tissue, and placenta in infected pregnant WT and IFNGR1-KO mice (Fig 3C), while the distribution of luciferase activity in organs from infected pregnant IFNGR1-KO mice was similar to that in infected pregnant WT mice on day 6 p.i. (Fig 3C and  3D). These results suggest that the accumulation of erythrocytes infected with PbNK65L was not affected by IFNGR1 deficiency.

IFNGR1 signaling plays a crucial role in placental inflammation during PbNK65L infection
We then performed histological analyses of placenta (Fig 4). As a result, a decreasing number of vascular branches and accumulation of infected erythrocytes were observed in the labyrinth of placentas from infected pregnant WT mice, but not in uninfected pregnant WT mice ( Fig  4A and 4B). However, accumulation of infected erythrocytes within fetal vascular were not observed in the placentas from infected pregnant WT mice (Fig 4C and 4D). In contrast, the decreasing numbers of vascular branches were improved in the labyrinth of placentas from infected pregnant IFNGR1-KO mice (Fig 4A and 4B). On the other hand, accumulation of infected erythrocytes within maternal vasculature, but not within fetal vasculature, was observed (Fig 4C and 4D). These results suggested that IFNGR1 signaling is involved in placental inflammation during PbNK65L infection.

Accumulation of CD8 + T cells and F4/80 + cells in the placenta is suppressed in IFNGR1-KO mice
During pregnancy, in the labyrinths of placentas from infected WT mice, the intervillous space and fetal weight were significantly decreased compared with those in uninfected WT mice on day 6 infection (Fig 5A and 5B). The decreased intervillous space and fetal weight were improved in infected IFNGR1-KO mice (Fig 5A and 5B). To investigate the effect of IFNGR1-deficiency on the immune response in the placenta during malaria infection, CD4 + T, CD8 + T, and F4/80 + cells in the placenta were assessed by flow cytometry (Fig 5C-5E and S2  Fig). The proportion of CD4 + T cells in placentas from infected WT mice was lower than that in uninfected mice (Fig 5C), while the proportions of CD8 + T and F4/80 + cells in placentas from infected WT mice were higher than those in uninfected mice (Fig 5D and 5E). The decreased proportion of CD4 + T cells and increased proportions of CD8 + T and F4/80 + cells were recovered in IFNGR1-deficiency (Fig 5C-5E). These findings suggest that IFNGR1 signaling is involved in the increases in CD8 + T and F4/80 + cells in the placenta. Therefore, we examined the effect of CD8 + T cell and neutrophil/macrophage depletion on pregnancy outcomes during malaria infection. We observed adverse pregnancy outcomes during malaria infection in CD8 + T cell-depleted or neutrophil/macrophage-depleted mice (Table 1). Our findings suggest that CD8 + T and F4/80 + cells might not be associated with the development of placental pathology during malaria.

Discussion
This study explored the role of IFNGR1 in the pathogenesis of placental pathology using a mouse model of malaria during pregnancy. The bioluminescence imaging results suggested that the accumulation of PbNK65L-infected erythrocytes in organs such as the lung, spleen, subcutaneous adipose tissue, and placenta in infected pregnant IFNGR1-KO mice was comparable to that in infected pregnant WT mice. The PbNK65L-infected pregnant WT mice showed adverse pregnancy outcomes. Here, we found that fetal mortality in PbNK65L-infected IFNGR1-KO mother mice was significantly lower than in WT mice infected with PbNK65L. Histological analyses of placentas showed that damage to the chorionic villi in PbNK65L-infected IFNGR1KO mice was improved compared with that in infected WT mice. These findings suggest that the accumulation of infected erythrocyte to placenta could induce inflammatory responses via IFNGR1 signaling, and are associated with adverse pregnancy outcomes during infection with malaria parasites. Numerous studies have shown that placental inflammation is induced in women and mice infected with malaria parasites [16][17][18]. Malaria parasites are recognized by pathogen-associated molecular patterns, such as toll-like receptors (TLRs), of host immune cells [39]. In addition to immune cells, the expression of the TLR family was observed in trophoblasts and syncytiotrophoblasts [40,41]. In a mouse model, P. berghei NK65-infected erythrocytes adhered to placental tissue by binding to CSA and were involved in the induction of pro-inflammatory cytokines in placental tissue [30]. High levels of pro-inflammatory cytokines in the placenta have been shown to be abrogated by deficiency in MyD88 during malarial infection [30], considering that proinflammatory cytokines such as IFN-γ may be produced by host immune cells and trophoblasts when TLR signaling is activated by the accumulation of infected erythrocytes in the placenta.
In this study, the parasite accumulation in the placentas of pregnant IFNGR1-deficient mice was comparable to that in WT mice infected with PbNK65L. However, the proportions of CD8 + T and F4/80 + cells in the placentas of infected IFNGR1-deficient mice were lower than those in infected WT mice. These results suggest that IFNGR1 signaling is involved in the increase in immune cells, such as CD8 + cells, neutrophils, and macrophages, in placental tissue. IFN-γ is associated with fetal abortion and resorption during Plasmodium and bacterial infection [19,42]. The infiltration of neutrophils and macrophages is observed in placentas from patients infected with malaria parasites [13,14]. During bacterial infection, CD8 + T cells and neutrophils are related to adverse pregnancy outcomes [42]. Although mice were depleted of CD8 + cells or neutrophil/macrophages, preterm delivery and the pup survival rate were not improved in CD8 + cell-or neutrophil/macrophage-depleted mice (Table 1). These findings suggest that both CD8 + T cells and F4/80 + cells might not be involved in the development of placental pathology and the mechanism by which placental pathology is developed during Plasmodium infection may be more complicated than in bacterial infection.
TNF-α is one of the major macrophage-produced cytokines and is associated with the development of placental pathology during malaria [31]. Although IFN-γ is a major inducer of TNF production, Poovassery et al. [31] have shown that pregnant IFN-γ deficient mice infected with P. chabaudi AS exhibited high levels of TNF compared to uninfected pregnant mice. For instanse, high levels of TNF-α activity were detected when monocytes were cocultured with P. falciparum schizont stage-parasitized erythrocytes or pigment recovered from ruptured schizonts [43,44]. The previous study by Poovassery et al. [31] suggest that IFNGR1 signaling may affect the expression level of TNF receptor or number of TNF receptor expressed cells but not affect TNF production during malaria. IFN-γ production is restricted to cells of the immune system [45]. However, because IFNGR1/2 proteins are widely expressed, nearly all cell types are capable of responding to IFN-γ [45]. Moreover, it is reported that IFNGR1 is also localized in the placenta throughout pregnancy [46,47]. It remains unclear which cells are important for IFN-γ signaling in the development of placental pathology during malaria.
We observed high levels of luciferase activity in adipose tissue in infected pregnant mice on day 3 p.i. A study using P. berghei ANKA showed that schizont membrane-associated cytoadherence protein (SMAC) is involved in the CD36-mediated sequestration of schizonts in adipose tissue [48]. The accumulation of erythrocytes infected with PbNK65L in subcutaneous adipose tissue in pregnant mice may be involved in the CD36-mediated sequestration via SMAC. During pregnancy, lipid metabolism is drastically changed for fetal development through the action of pregnancy hormones, such as progesterone [49] or human chorionic gonadotropin [50]. Therefore, the accumulation of infected erythrocytes into adipose tissue may be due to enhanced expression of CD36 on adipose tissue or increases in CD36-expressing adipocytes during pregnancy. Our findings suggest that the accumulation of infected erythrocytes in adipose tissue increases the sensitivity and specificity of detecting infection during pregnancy.
A rapid increase of parasitemia in pregnant mice was observed compared with that in nonpregnant mice. Regulatory T cells have been shown to increase during pregnancy [51]. Therefore, it is believed that the increase in susceptibility of mice to infection with malaria parasites may be associated with regulatory T cells during pregnancy. Pregnancy-associated hormones, such as progesterone (P4) and 17β-estradiol (E2), are known to enhance the suppressive capacity of regulatory T cells [52]. Although no effect of treatment of P4 and/or E2 on the course of parasitemia was observed in mice infected with P. chabaudi, treatment with E2 increased levels of IL-10 in the infected mice [53]. In fact, high levels of IL-10 were observed in pregnant mice in this study. On the other hand, maternal metabolism, such as glucose, lipid [54,55], and purine metabolism [56], changes significantly during pregnancy considering that the alteration in host nutrient conditions may be associated with an enhancement of parasite-growth during pregnancy.
Our finding that IFNGR1 signaling plays a central role in the development of placental pathology during malaria in pregnancy suggests that blocking IFNGR1 signaling may be effective for the prevention of adverse pregnancy outcomes during infection with malaria parasites. In addition, our findings suggest that the accumulation of infected erythrocytes in adipose tissue in pregnant mice during the early phase of infection may contribute to the identification of a parasite marker for early detection of malaria parasites during pregnancy. Based on these preclinical findings, additional investigations are required to establish whether maternal IFNGR1 signaling is also involved in developing placental pathology in clinical cases of severe malaria. Restriction sites of NheI and BglII restriction enzymes were shown. (C) Luciferase (luc2)expressing cassette was introduced into target gene by double-crossover homologous recombination. Arrows (F1 and F2) denote primers specific for the 5 0 and 3 0 regions of the target gene (S1 Table). Introduction of luc2 into the p230 locus (PBANKA_030600), which is not essential in the complete life cycle of the parasite [36], of PbNK65 parasites. Proper integration was confirmed using primers specific for p230 (WT, 4.7 kbp; Introduced, 7.0 kbp) for two cloned transfected parasites. Parasites in which the luc2-expressing cassette was introduced into the p230 locus were used as control PbNK65L in this study.  Table. The effect of different timing of infection on the pregnancy outcomes. a; Mice on day 12, or 15 post-mating were infected with 1 × 10 4 erythrocytes infected with luciferaseexpressing P. berghei NK65 (PbNK65L). Experiments using three mice were performed in duplicate with similar results. P values indicate a significant difference compared to pregnancy period of uninfected mice. (XLSX)